The automatization of social media communication - Exploring the development of bot codes on GitHub and the use of open-source bots on Twitter

Kivonat, rövid leírás

The production and deployment of social media bots signal the emergence of new political economies that redistribute agency around new technological actors. This has implications for marketing, political action and even private lives. Yet, we have very little systematic knowledge about how bots are produced, and how open-source developers use collaborative platforms for sharing their code online. In my PhD research project, I proposed a novel way to study Twitter bots by examining both the open-source bot codes on GitHub and the bots deployed on Twitter. Traditionally, research projects have examined bots as black boxes and attempted to “reverse engineer” the algorithms behind automated accounts. Instead, I tracked bots to their source code on GitHub and connect the algorithm (code) running behind the bot to data about bot activity on a social media platform. By using this approach, I combined data about automated accounts deployed on a social media platform with the source code behind those same bots – this creates a unique lens on bots and provides important insights about how they were produced and how they operate. Combining API-based, data driven research with a classic social science approach helps to understand current practices of developing bots and deploying them on online social media platforms. As part of my PhD research, I examined bot repositories on GitHub and conducted a survey with the same bot developers to gain further insights into the nitty-gritty of writing bot code, including the most important challenges during the development phase and major barriers to deploying and operating bots on Twitter over an extensive period of time. The survey results also shed light on the motivations behind creating automated social media accounts. I also studied how programmers deploy their bots on Twitter and how other, mostly human, users react to the bot activity. The final part of my thesis contributes to a typology of open-source social media bots by systematically examining bot codes shared on GitHub and the activity of some of these bots on Twitter in tandem. My dissertation contributes to scientific knowledge about open source Twitter bots in two distinct ways. First, it provides a framework to study both the development and the use of open source bots on Twitter. Secondly, it explores the practices of developing open source bots and shed lights on the motivation behind developing Twitter bots, the long-term use of such bots including the challenges of running bots for an extensive period of time and the bots contribution to the Twitter eco-system.